Feature selection algorithms

To enable the algorithms to train faster, and to reduce the complexity and overfitting of the model, in addition to improving its accuracy, you can use many feature selection algorithms and techniques. We are going to look at three different feature selection methods: filter methods, wrapper methods, and embedded methods. Let's discuss the various methodologies and techniques.

Filter methods

In filter methods, each feature will be assigned a score, computed by different statistical measures. In other words, these methods rank features by considering the relationships between the features and the targets. Filter methods are usually used in the pre-processing phase:

Pearson's correlation

Pearson's correlation is a statistical method used to measure the linear correlation between two variables, x and y. It is ranged between +1 and -1 ; +1 means that there is a positive association. You need to know that x and y should be continuous variables. The formula for Pearson...